Energy use has been a big problem in the data centre for, some might say, as
much as five years. Throughout 2008, it has been a headline issue: severe
power shortages and prices are, in some geographies, limiting expansion and
investment; suppliers regularly claim product breakthroughs and launch new
services; and some data centre operators and end user organisations have
reported major successes.By Andy Lawrence, Research Director,
Eco-Efficient IT, The 451 Group.
How data like this is used is the focus of a big debate. In Overhaul data centre cooling. This is the low-hanging fruit
the past year, the PUE (power usage effectiveness) ratio, of data centre energy efficiency. It involves multiple
which measures the ratio of total facility power against the strategies: separating hot and cold air; the use of advanced
total use of IT power, has become the de facto metric for technologies such as targeted and in-row cooling; raising
measuring data centre efficiency (along with its reciprocal, chiller temperatures and increasing hot aisle temperatures,
DCiE). It is, or will be, adopted by, for example, Energy Star, and upgrading to energy efficient motors and pumps.
and partially, by the European Commission.
Use free air cooling. The use of free air cooling, using
That is a good thing – but this figure tends to focus minds economisers, is cited by many of the most high energy
on the efficiency of the cooling, rather than on the efficiency efficient data centres.
of the IT– and it says nothing about efficiency over time, the
location or specific tasks of the data centre, or the type of These examples of best practice do not mean that the most
energy being used. For this reason, parallel initiatives are energy efficient data centres will use all of these. There are
underway to measure the “useful work” of the IT load some signs that the most efficient data centres fall into one
(Green Grid) or the total facility efficiency (Uptime Institute). of two groups – high density systems with in-row or
To date, these measurements have not gained traction, and dynamic, targeted cooling; or lower density data centres
it is not yet easy to truly assess what makes one data with free-air cooling.
centre greener than.
What about those that have not worked so well? These fall
Best technologies and practices into two categories – those that don’t make much
Even so, it is clear that some of the many technologies and difference, such as some energy efficiency forms of storage;
methodologies for the green data centre work well – and and those that are not yet been proven or adopted, or may
others do not work so well. prove too expensive.
It is controversial territory, but at The 451 Group, we believe The latter category includes container data centres, server
the following technologies have worked well: data centre power management (turning machines off),
water cooling (effective, but not widely used), relocation to
Consolidation and virtualisation. Even if the project might low density or cold climates; and switching to DC power.
have happened anyway (and even if power savings are
usually less than the 50% often claimed). Finally, in recent months, some big data centre operators,
such as Microsoft and Google, have begun publishing
Energy efficient (blade) servers. Multi-core, energy efficient detailed PUE and other figures about their data centres,
servers can bring energy use down because they replace joining companies such as BT and others. In each case,
and consolidate many older servers: they are far more dramatic PUE figures have been claimed (in one Google
energy efficient (transactions or mips per watt) than older data centre, just 1.13), raising all kinds of questions. But
models; and any savings in IT equipment will result in what these examples really reveal is that flagship data
upstream or downstream savings in UPS or cooling power centre operators have accepted the challenge of being as
(assuming variable control equipment). energy efficient as they possibly can be.
15
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73 |
Page 74 |
Page 75 |
Page 76